Multi-Objective Synthesis of CMOS Operational Amplifiers using a Hybrid Genetic Algorithm


Abstract

The design of an analogue integrated circuit is a complex and tedious task which requires many compromises to be made between conflicting objectives. This work focuses on the synthesis and design of CMOS operational amplifiers (opamps). We show how a multi-objective Genetic Algorithm (GA) [5] is used to search the design where many of the objectives that need to be satisfy are conflicting. In this paper, we introduce a novel coding scheme where both the structure and value parameters of the circuit can be encoded in a single chromosome. We further show how this coding scheme can be translated into a real circuit for a Pareto-based multi-objective algorithm to solve the problem.